肺癌短期重复扫描ADC值和IVIM参数值的测量可重复性

目的:前瞻性评估肺癌短期重复扫描的ADC值及体素内不相干运动(IVIM)参数值的测量可重复性.方法:27个肺癌患者(27个病灶)进行2次自由呼吸DWI(b=0,300,800 s/mm2)及IVIM(10个b值,b=0～1000 s/mm2)扫描(间隔30～60min).2名独立测量者分别在ADC图、IVIM图上通过勾勒病灶获得ADC均值和D、D*、f值.采用配对样本t检验(正态分布)或Wilcoxon检验(非正态分布)比较重复测量、不同测量者和重复扫描的误差.各参数值的测量者内可重复性、测量者间一致性以及两次扫描的测量可重复性采用组内相关系数(ICC)、组间变异系数(WCV)和Bland-Altman法评价.结果:重复测量和重复扫描所得ADC均值、IVIM参数值差异均无统计学意义(P＞0.05).所有参数显示出好的测量者内可重复性和测量者间一致性,除了D*(WCV＞30％).重复扫描D值的可重复性最佳(ICC为0.905、0.883,WCV为7.63％、8.46％),ADC值的可重复性较好(ICC为0.829、0.823,WCV为10.65％、11.13％),而D*和f的可重复性相对较差(ICC为0.639～0.802;WCV为39.50％～58.14％).结论:当运用DWI或IVIM来监测肺癌疗效时,若ADC值、D、D*和f值的变化分别小于21.81％、16.58％、113.95％和84.77％,则很可能是由测量误差引起.需要改进IVIM扫描技术和后处理算法,提高D*和f的测量可重复性.

Short-term reproducibility of apparent diffusion coefficient and intravoxel incoherent motion parameters in lung cancer patients

Objective:To prospectively evaluate the short-term test-retest reproducibility of diffusion-weighted imaging (DWI) and Intravoxel Incoherent Motion (IVIM) parameters in lung cancer patients.Methods:27 patients with lung cancer (27 lesions) underwent free-breathing DWI (b=0,300,800s/mm2) and IVIM (10 b-values,0 ～ 1000s/mm2) scans twice (30～60m interval).Regions of interests were drawn on apparent diffusion coefficient (ADC) maps and IVIM images to derive ADCs and IVIM parameters D,D* and f by two independent observers.All the parameters were compared between repeated measurements,different observers and repeated scans using the paired-sample t-test (normality) or Wilcoxon signed-rank test (non-normality).Intra-and inter-observer,test-retest reproducibility and variability was assessed with interclass correlation coefficient (ICC),within coefficient of variation (WCV),and Bland-Altman analysis.Results:There were no significant difference between repeated measurements and repeated scans for both mean ADCs and IVIM parameters (P＞0.05).Apart from D* (WCV＞ 30 ％),all the parameters showed good intra-and inter-observer agreement.The test-retest reproducibility of D was the best (ICC 0.905,0.883;WCV 7.63 ％,8.46 ％),ADC was good (ICC 0.829,0.823;WCV 10.65％,11.13％),while D* and f was poor (ICC 0.639～0.802;WCV 39.50M～58.14％ respectively).Conclusion:Using DWI or IVIM to monitor the treatment effect of lung cancer,if the change of ADC,D,D* and f is less than 21.81％,16.58％,113.95％ and 84.77 ％ respectively,it might be caused by measurement error.Improvement of IVIM scanning technology and post-processing algorithms are needed to heighten the measurement repeatability of D* and f.